Lexer is significantly better tested and more advanced

This commit is contained in:
Aidan Hahn 2019-11-28 23:51:19 -08:00
parent 65cecb3647
commit aaa0a8a507
No known key found for this signature in database
GPG key ID: 327711E983899316
3 changed files with 47 additions and 16 deletions

Binary file not shown.

BIN
shsh Executable file

Binary file not shown.

View file

@ -31,9 +31,12 @@ func Lex(input string) *Token {
var tok strings.Builder var tok strings.Builder
is_list := false is_list := false
is_str := false is_str := false
is_num := true
tokenBuilder := func (pos int, tok string) { tokenBuilder := func (pos int, tok string) {
if len(tok) == 0 && !is_list && !is_str {
return
}
*iter = new(Token) *iter = new(Token)
(*iter).position = pos (*iter).position = pos
@ -48,7 +51,7 @@ func Lex(input string) *Token {
(*iter).tag = STRING (*iter).tag = STRING
is_str = false is_str = false
} else if is_num { } else if tokenIsNumber(tok) {
(*iter).tag = NUMBER (*iter).tag = NUMBER
} else { } else {
@ -60,33 +63,61 @@ func Lex(input string) *Token {
} }
for pos, char := range input { for pos, char := range input {
switch char { if char == delim {
// TODO: User configurable delimiters if is_str && is_list {
case '\'', '"', '`': // String just ended inside list
is_str = true is_str = false
delim = char delim = ')'
case '(': tok.WriteRune(char)
is_list = true
delim = ')'
case delim:
delim = ' '
if tok.Len() == 0 {
continue continue
} }
delim = ' '
tokenBuilder(pos, tok.String()) tokenBuilder(pos, tok.String())
tok.Reset() tok.Reset()
default: } else {
is_num = is_num && unicode.IsDigit(char) if strings.ContainsRune("\"'`", char) {
is_str = true
delim = char
if !is_list {
continue
}
} else if char == '(' && !is_str {
is_list = true
delim = ')'
continue
}
tok.WriteRune(char) tok.WriteRune(char)
} }
} }
if tok.Len() > 0 { if tok.Len() > 0 {
if is_list || is_str {
// TODO: Throw hella lex error here
return ret
}
tokenBuilder(len(input), tok.String()) tokenBuilder(len(input), tok.String())
} }
return ret return ret
} }
func tokenIsNumber(arg string) bool {
dotCount := 0
for _, char := range arg {
if !unicode.IsDigit(char) {
if char == '.' && dotCount == 0 {
dotCount++
} else {
return false
}
}
}
return true
}